arXiv subjects

Harsh Patel

Publications and source records attributed to Harsh Patel.

22 records · Page 2Linked to original sources

arXiv2021

HinFlair: pre-trained contextual string embeddings for pos tagging and text classification in the Hindi language

Recent advancements in language models based on recurrent neural networks and transformers architecture have achieved state-of-the-art results on a wide range of natural language processing tasks such as pos tagging, named entity recognition, and text classification. However, most of these language models are pre-trained in high resource languages like English, German, Spanish. Multi-lingual language models include Indian languages like Hindi, Telugu, Bengali in their training corpus, but they often fail to represent the linguistic features of these languages as they are not the primary language of the study. We introduce HinFlair, which is a language representation model (contextual string embeddings) pre-trained on a large monolingual Hindi corpus. Experiments were conducted on 6 text classification datasets and a Hindi dependency treebank to analyze the performance of these contextualized string embeddings for the Hindi language. Results show that HinFlair outperforms previous state-of-the-art publicly available pre-trained embeddings for downstream tasks like text classification and pos tagging. Also, HinFlair when combined with FastText embeddings outperforms many transformers-based language models trained particularly for the Hindi language.

cs.CL↗

arXiv2020

BioNerFlair: biomedical named entity recognition using flair embedding and sequence tagger

Motivation: The proliferation of Biomedical research articles has made the task of information retrieval more important than ever. Scientists and Researchers are having difficulty in finding articles that contain information relevant to them. Proper extraction of biomedical entities like Disease, Drug/chem, Species, Gene/protein, can considerably improve the filtering of articles resulting in better extraction of relevant information. Performance on BioNer benchmarks has progressively improved because of progression in transformers-based models like BERT, XLNet, OpenAI, GPT2, etc. These models give excellent results; however, they are computationally expensive and we can achieve better scores for domain-specific tasks using other contextual string-based models and LSTM-CRF based sequence tagger. Results: We introduce BioNerFlair, a method to train models for biomedical named entity recognition using Flair plus GloVe embeddings and Bidirectional LSTM-CRF based sequence tagger. With almost the same generic architecture widely used for named entity recognition, BioNerFlair outperforms previous state-of-the-art models. I performed experiments on 8 benchmarks datasets for biomedical named entity recognition. Compared to current state-of-the-art models, BioNerFlair achieves the best F1-score of 90.17 beyond 84.72 on the BioCreative II gene mention (BC2GM) corpus, best F1-score of 94.03 beyond 92.36 on the BioCreative IV chemical and drug (BC4CHEMD) corpus, best F1-score of 88.73 beyond 78.58 on the JNLPBA corpus, best F1-score of 91.1 beyond 89.71 on the NCBI disease corpus, best F1-score of 85.48 beyond 78.98 on the Species-800 corpus, while near best results was observed on BC5CDR-chem, BC3CDR-disease, and LINNAEUS corpus.

cs.CL↗

arXiv2020

Assessing the Interplay between travel patterns and SARS-CoV-2 outbreak in realistic urban setting

The dense social contact networks and high mobility in congested urban areas facilitate the rapid transmission of infectious diseases. Typical mechanistic epidemiological models are either based on uniform mixing with ad-hoc contact processes or need real-time or archived population mobility data to simulate the social networks. However, the rapid and global transmission of the novel coronavirus (SARS-CoV-2) has led to unprecedented lockdowns at global and regional scales, leaving the archived datasets to limited use. While it is often hypothesized that population density is a significant driver in disease propagation, the disparate disease trajectories and infection rates exhibited by the different cities with comparable densities require a high-resolution description of the disease and its drivers. In this study, we explore the impact of the creation of containment zones on travel patterns within the city. Further, we use a dynamical network-based infectious disease model to understand the key drivers of disease spread at sub-kilometer scales demonstrated in the city of Ahmedabad, India, which has been classified as a SARS-CoV-2 hotspot. We find that in addition to the contact network and population density, road connectivity patterns and ease of transit are strongly correlated with the rate of transmission of the disease. Given the limited access to real-time traffic data during lockdowns, we generate road connectivity networks using open-source imageries and travel patterns from open-source surveys and government reports. Within the proposed framework, we then analyze the relative merits of social distancing, enforced lockdowns, and enhanced testing and quarantining mitigating the disease spread.

physics.soc-ph↗

arXiv2019

Split-Scale: Scaling Bitcoin by Partitioning the UTXO Space

The Bitcoin protocol is a significant milestone in the history of money. However, its adoption is currently constrained by the transaction limits of the system. As the chief problem of blockchain technology, the scaling issue has attracted many valuable solutions both on-chain and off-chain. In this paper, our goal is to explore the notion of unspent transaction outputs (UTXOs) to propose an augmented Bitcoin protocol that can scale gracefully. Our proposal aims to increase the transaction throughput by partitioning the UTXO space and splitting the blockchain. In addition, a new type of Bitcoin node is introduced to preserve the capability to run validating nodes in low-bandwidth environments, despite the increased transaction throughput.

cs.CR↗